DSPy

DSPy calls models through LiteLLM, so OpenCheap is an openai/ model with an api_base.

What you need

  • Base URLhttps://api.opencheap.si/v1
  • API keyA kl_live_… key from the console, kept in OPENCHEAP_API_KEY.
  • ModelAn id from Models, for example claude-sonnet-5.
pip install dspy

The language model

import os
import dspy

lm = dspy.LM(
    "openai/claude-sonnet-5",
    api_base="https://api.opencheap.si/v1",
    api_key=os.environ["OPENCHEAP_API_KEY"],
)
dspy.configure(lm=lm)

qa = dspy.Predict("question -> answer")
print(qa(question="What is the capital of France?").answer)

cot = dspy.ChainOfThought("question -> answer: int")
print(cot(question="What is 6 times 7?").answer)

openai/ is LiteLLM's prefix for an OpenAI-compatible endpoint; claude-sonnet-5 is what OpenCheap receives. Every module you build after dspy.configure uses this model, and an optimizer's calls are billed to your key like any other.

What was checked

Run on October 28, 2026 with DSPy 3.4.1 on claude-sonnet-5: the program above answered "Paris" and 42. Optimizers were not run.